Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SYNTHESIS OF DEOXYAMINO SUGARS
Document Type and Number:
WIPO Patent Application WO/2014/128660
Kind Code:
A1
Abstract:
The invention relates to methods for making an α-amino-aldehyde, preferably a 2-amino-2-deoxy sugar derivative, from an α-hydroxy-aldehyde, preferably an aldose. Intermediary 1-amino-1-deoxy-ketose derivatives, and their use as synthetic intermediates, are also described.

Inventors:
KHANZHIN NIKOLAY (DK)
CHASSAGNE PIERRE (FR)
MATWIEJUK MARTIN (DE)
HORVÁTH FERENC (HU)
DEKANY GYULA (AU)
Application Number:
PCT/IB2014/059159
Publication Date:
August 28, 2014
Filing Date:
February 21, 2014
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
GLYCOM AS (DK)
International Classes:
C07H5/06; C07H15/12
Domestic Patent References:
WO2007104311A12007-09-20
WO2009059945A22009-05-14
Foreign References:
AT503400A12007-10-15
Other References:
WRODNIGG T M ET AL.: "The Amadori and Heyns rearrangements: Landmarks in the history of carbohydrate chemistry or unrecognized synthetic opportunities?", TOPICS IN CURRENT CHEMISTRY, vol. 215, 2001, pages 115 - 152
WRODNIGG TM ET AL.: "The Heyns rearrangement revisited:an exceptionally simple two-step chemical synthesis of D- Lactosamine from Lactulose", COMMUNICATIONS, ANGEW. CHEM.INT.ED., vol. 38, no. 6, 1999, pages 827 - 828, XP002431802, DOI: doi:10.1002/(SICI)1521-3773(19990315)38:6<827::AID-ANIE827>3.0.CO;2-N
SHAN Y.: "Lactosamine from lactulose via the Heyns rearrangement: a practical protocol", TETRAHEDRON LETTERS, vol. 54, 2013, pages 3960 - 3961, XP028573303, DOI: doi:10.1016/j.tetlet.2013.05.086
WRODNIGG T M ET AL.: "The Amadori rearrangement as key reaction for the synthesis of neoglycoconjugates", CARBOHYDRATE RESEARCH, vol. 343, 2008, pages 2057 - 2066, XP022757842, DOI: doi:10.1016/j.carres.2008.02.022
LEVI J ET AL.: "Fluorescent fructuse derivatives for imaging breast cancer cells", BIOCONJUGATE CHEMISTRY, vol. 18, 2007, pages 628 - 634, XP008142401, DOI: doi:10.1021/bc060184s
Attorney, Agent or Firm:
SELDEN, Deborah, A. (Fulwood House12 Fulwood Place, London London WC1V 6HR, GB)
Download PDF:
Claims:
A method for making an a-amino-aldehyde, wherein the method comprises either:

(A) the step of reacting an a-hydroxy-aldehyde with at least two equivalents of an amine reagent in the presence of an acid catalyst; or alternatively

(B) the steps of:

a) reacting an a-hydroxy-aldehyde with an amine reagent to obtain an a-amino-ketone, and b) reacting the a-amino-ketone obtained in step a) with an amine reagent, preferably with the same amine reagent used in step a), to obtain an a-amino-aldehyde.

The method according to claim 1, wherein the amine reagent is represented by the formula HN-R1 2, wherein Ri and R2 are, independently, H, alkyl, homoaryl, heteroaryl or benzyl, or Ri and R2 together form a divalent -(CH2)n- moiety wherein n is 4-8 and wherein a methylene group can be replaced by an oxygen, a sulphur or a > -R3 moiety wherein R3 means H or alkyl.

The method according to claim 2, wherein the amine reagent is a primary amine selected from the group of optionally substituted alkyl, homoaryl, heteroaryl and benzyl amine, preferably optionally substituted benzyl amine, more preferably benzyl amine.

The method according to any of claims 1 to 3, wherein the α-hydroxy-aldehyde is a mono-, di- or oligosaccharide aldose of formula 1, and the α-amino-aldehyde is a mono-, di- or oligosaccharide 2-amino-2-deoxy-aldose of formula 2a or 2b

wherein Ri and R2 are H, optionally substituted alkyl, homoaryl, heteroaryl or benzyl, or Ri and R2 together form a divalent -(CH2)n- moiety wherein n is 4-8 and a methylene group can be replaced by an oxygen, a sulphur or a >N-R3 moiety wherein 3 means H or alkyl, R4, independently, is H or -OR5 wherein R6 is H or a sugar component, and R5 is H, -COOH or -CH2R4.

5. The method according to claim 4, wherein a compound of formula 1 is transformed to a compound of formula 2a via a compound of formula 2b.

6. The method according to claim 4 or 5, wherein compounds of formulae 1, 2a and 2b are monosaccharides or disaccharides

7. The method according any one of claims 1 to 6, wherein the a-hydroxy-aldehyde is glucose or lactose, and the amine reagent is benzyl amine.

8. The method according to any of the claims 1 to 7, wherein the method comprises alternative

(B).

9. The method according to claim 8, wherein steps a) and b) are each carried out in the

presence of a catalyst, preferably acetic acid in step a) and an acid addition salt of the amine reagent, more preferably the hydrochloride salt of the amine reagent, in step b).

10. The method according to claims 8 or 9, wherein the a-amino-ketone is a mono-, di- or oligosaccharide ketose of formula 3

3

wherein Ri and R2 are H, optionally substituted alkyl, homoaryl, heteroaryl or benzyl, or Ri and R2 together form a divalent -(CH2)n- moiety wherein n is 4-8 and wherein a methylene group can be replaced by an oxygen, a sulphur or a > -R3 moiety wherein R3 means H or alkyl; wherein each R4, independently, is H or -ORe wherein 5 is H or a sugar component; and wherein R5 is H, -COOH or -CH2R4.

11. The method according to claim 10, wherein a compound of formula 3 is transformed to a compound of formula 2a via a compound of formula 2b.

12. The method according to claim 10 or 11, wherein compounds of formula 3 are

monosaccharides or disaccharides.

13. The method according any one of claims 8 to 12, wherein the α-hydroxy-aldehyde is glucose or lactose, and the amine reagent in both steps a) and b) is benzylamine.

14. The method according to any one of claims 1 to 7, wherein the method comprises alternative (A), and the catalyst is an acid addition salt of the amine reagent, more preferably the hydrochloride salt of the amine reagent.

15. A compound of formula 6

wherein Ria and R.2a are each H or a group removable by hydrogenolysis, R4a, independently, is H or -0¾3, Rsa is H, -COOH or -CH2R4a, and R6a is H or a sugar component, provided that at least one of Ria or R.2a is a group removable by hydrogenolysis and provided that there is one ¾α in ILta and Rsa that is a sugar component and the others are H,

or a salt thereof.

16. The compound according to claim 15 characterized by formula 6a

wherein Ria, R.2a, R4a and Rsa are as defined above.

17. The compound according to claim 16, which is 1-benzylamino-l-deoxy-lactulose or a salt thereof.

18. Use of a compound of any one of claims 15 to 17 as a synthetic intermediate.

Description:
SYNTHESIS OF DEOXYAMINO SUGARS

FIELD OF THE INVENTION

The present invention relates to the field of carbohydrate chemistry, namely the synthesis of 2-amino-2-deoxy sugar derivatives from aldoses.

BACKGROUND OF THE INVENTION

A sugar in which one or more nonglycosidic hydroxyl groups is replaced by an amino or substituted amino group is called an amino or deoxyamino sugar. The most abundant example is D- glucosamine (2-amino-2-deoxy-D-glucose). D-Glucosamine, mostly in N-acetylated form

(GlcNAc), is a part of the cell wall oligosaccharides of bacteria, fungi and other organisms. GlcNAc is the monomeric unit of chitin. 4-O-Galactosylated GlcNAc (N-acetyllactosamine) is one of the most common components of natural oligosaccharides like Le x -type oligosaccharides or human milk oligosaccharides which play essential roles in important biological events. Lacto-N-biose (3- O-galactosyl GlcNAc) is a fundamental part of the family of Le a -type oligosaccharides, Type 1 human ABH antigenic determinants. Lacto-N-biose containing oligosaccharides like lacto-N- tetraose and its fucosylated and/or sialylated derivatives are among some of the major components in human milk. 2-Amino-2-deoxy-D-mannose (D-mannosamine) mainly in its N-acetylated form (ManNAc) can be found as a building unit of some bacterial capsular polysaccharides and lipopolysaccharides. In addition, N-acetyl-D-mannosamine is the biosynthetic precursor of sialic acid, a unique nine-carbon ketoaldonic acid having many maj or biological roles.

In classical synthetic methods, cheap and easily available simple monosaccharides or disaccharides can be transformed into their respective amino sugar framework. These

methodologies involve manipulations at C-2 (e.g nucleophilic displacement of good leaving groups in C-2 with N-nucleophiles, stereoselective reduction of 2-ulose oximes, intermolecular addition of nitrosyl chloride to the double bond of a glycal, azidonitration of the double bond of a glycal) in order to introduce or exchange the OH-group with amino, protected amino or masked amino (e.g. azido) function. These chemical pathways always stand in need of extensive use of protecting groups in order to mask functional groups that would be affected by the key transformations, thus they consist of many elementary chemical steps. Such multistep sequences restrict usefulness and are not attractive for large scale developments because of the long technological time and the use of high number of reagents (which in fact, not uncommonly, can be toxic, of low availability and/or expensive) and/or require lengthy or cumbersome isolation/separation procedures.

The reaction of a ketose with an amine giving a ketosyl amine and the subsequent rearrangement of the latter into 2-amino-2-deoxy-aldose is known as Heyns rearrangement (see Hodge Adv. Carbohydr. Chem. 10, 169 (1955), Wrodnigg et al. Top. Curr. Chem. 215, 1 15 (2001)). Theoretically, both 2-epimers can be formed, but the formation of one of the epimers is favoured, presumably because of steric factors. Depending on the reactions conditions, 2-amino-2-deoxy- aldoses (e.g. Carson J. Am. Chem. Soc. 77, 5957 (1955) and 78, 3728 (1956), Wrodnigg et al. Angew. Chem. Int. Ed. 38, 827 (1999), Stutz et al. J. Carbohydr. Chem. 22, 253 (2003), Shan et al. Tetrahedron Lett. 54, 3690 (2013), WO 94/26760, AT 503400 Al) or their N-glycosides (e.g. Heyns et al. Chem. Ber. 88, 1551 (1955), Piispanen et al. J. Org. Chem. 68, 628 (2003), WO 2009/059945, WO 2012/140576) can be obtained.

Kuhn et al. described (Chem. Ber. 87, 1547 (1954)) that catalytic hydrogenation of lactosazone led to a mixture of lactosamine (2-amino-2-deoxy-lactose) and isolactosamine (1- amino-l-deoxy-lactulose).

The reaction of an aldose with an amine giving an aldosyl amine and the subsequent rearrangement of the latter into 1 -amino- 1-deoxy-ketose is known as Amadori rearrangement (see Wrodnigg et al. Top. Curr. Chem. 215, 115 (2001)). Amadori-type rearrangement has also been observed in a non-sugar oc-hydroxy aldehyde reaction with an amine (see Itakura et al. Chemistry and Physics of Lipids 124, 81 (2003)).

Because of the biological importance of amino sugar derivatives, shorter and simpler methods have continuously been sought for synthesizing these derivatives. It is an aim of the present invention to provide such a method.

BRIEF DESCRIPTION OF THE INVENTION

The invention relates to a method for making an a-amino-aldehyde, preferably a 2-amino-2- deoxy sugar derivative, wherein the method comprises either:

the step of reacting an a-hydroxy-aldehyde, preferably an aldose, with at least two equivalents of an amine reagent in the presence of an acid catalyst; or alternatively

the steps of: a) reacting an -hydroxy-aldehyde, preferably an aldose, with an amine reagent to obtain an a-amino-ketone, preferably a 1 -amino- 1-deoxy-ketose derivative, and b) reacting the a-amino-ketone, preferably a 1 -amino- 1-deoxy-ketose derivative, obtained in step a) with an amine reagent, preferably with the same amine reagent used step a), to obtain an a-amino-aldehyde, preferably a 2-amino-2-deoxy sugar derivative.

In addition, the invention relates to providing compounds of the following formula 6

6 wherein Ri a and R 2a are H or a group removable by hydrogenolysis, each R4 a , independently, is H or -ΟΙ¾ 3 , s a is H, -COOH or -C¾R4a, and R6a is H or a sugar component, provided that at least one of Ri a or R 2a is a group removable by hydrogenolysis and provided that one R6 a in R4 a and Rs a is a sugar component and the others are H,

or a salt thereof.

DETAILED DESCRIPTION OF THE INVENTION

In accordance with this invention, the term "group removable by hydrogenolysis" preferably means a group whose bond attached to a core carbohydrate structure can be cleaved by addition of hydrogen in the presence of palladium, Raney nickel or another appropriate metal catalyst known for use in hydrogenolysis, resulting in the regeneration of the protected functional group, mainly -NH 2 of the parent molecule in the context of the present invention. Such protecting groups are well known and are thoroughly discussed in P.G.M. Wuts and T.W. Greene: Protective Groups in Organic Synthesis, John Wiley & Sons (2007). Suitable protecting groups include benzyl, diphenylmethyl (benzhydryl), 1 -naphthylmethyl, 2-naphthylmethyl and triphenylmethyl (trityl) groups, each of which can optionally be substituted by one or more groups selected from: alkyl, alkoxy, phenyl, amino, acylamino, alkylamino, dialkylamino, nitro, carboxyl, alkoxycarbonyl, carbamoyl, N-alkylcarbamoyl, N,N-dialkylcarbamoyl, azido, halogenalkyl and halogen. Preferably, such substitution, if present, is on the aromatic ring(s). Also in this invention, the term "alkyl" preferably means a linear or branched chain saturated hydrocarbon group with 1-10 carbon atoms, especially 1-6 carbon atoms, such as methyl, ethyl, ft-propyl, /-propyl, «-butyl, / ' -butyl, s-butyl, ί-butyl or «-hexyl; the term "homoaryl" preferably means a homoaromatic group such as phenyl or naphthyl; the term "heteroaryl" preferably means an aromatic group having one or two rings, which ring(s) contain(s) 1, 2, or 3 heteroatoms selected from the group of N, O and S, such as pyrrole, imidazole, pyrazole, 1,2,3- triazole, 1,2,4-triazole, furan, thiophene, oxazole, isoxazole, thiazole, thiadiazole, pyridine, pyrazine, pyridazine, pyrimidine, triazine, benzimidazole, benzoxazole, benzthiazole, indole, quinoline, isoquinoline, purine, pteridine and the like; the term "alkyloxy" or "alkoxy" preferably means an alkyl group attached to a parent molecular moiety through an oxygen atom, such as methoxy, ethoxy or /-butoxy; "halogen" means fluoro, chloro, bromo or iodo; "amino" means a - H 2 group; "alkylamino" preferably means an alkyl group attached to a parent molecular moiety through an - H-group, such as methylamino or ethylamino; "dialkylamino" preferably means two alkyl groups, either identical or different, attached to a parent molecular moiety through a nitrogen atom, such as dimethylamino or diethylamino; "acylamino" preferably means an acyl group attached to a parent molecular moiety through an -NH-group, such as acetylamino (acetamido) or benzoylamino (benzamido); "carboxyl" means a -COOH group; "alkyloxycarbonyl" preferably means an alkyloxy group attached to a parent molecular moiety through a -C(=0)-group, such as methoxycarbonyl or i-butoxycarbonyl; "carbamoyl" means an H 2 N-C(=0)-group; "N- alkylcarbamoyl" preferably means an alkyl group attached to a parent molecular moiety through a -HN-C(=0)-group, such as N-methylcarbamoyl; "N,N-dialkylcarbamoyl" preferably means two alkyl groups, either identical or different, attached to a parent molecular moiety through a >N-C(=0)-group, such as N,N-dimethylcarbamoyl.

With regard to the term "optionally substituted" in connection with the alkyl, homoaryl and heteroaryl groups, these groups can be substituted by one or more groups selected from: alkyl (only for homo- and heteroaryl residues), halogen, nitro, homoaryl, alkoxy, amino, alkylamino, dialkylamino, carboxyl, alkoxycarbonyl, carbamoyl, N-alkylcarbamoyl, N,N-dialkylcarbamoyl, azido, haloalkyl or hydroxyalkyl.

Also herein, the term "sugar component" preferably means a monosaccharide or oligosaccharide, particularly a monosaccharide. Monosaccharide preferably means a sugar of 5-9 carbon atoms that is an aldose (e.g. D-glucose, D-galactose, D-mannose, D-ribose, D-arabinose, L- arabinose, D-xylose, etc.), a ketose (e.g. D-fructose, D-sorbose, D-tagatose, etc.), a deoxysugar (e.g L-rhamnose, L-fucose, etc.), a deoxy-aminosugar (e.g. N-acetylglucosamine, N- acetylmannosamine, N-acetylgalactosamine, etc.), an uremic acid, a ketoaldonic acid (e.g. sialic acid) or equivalents. Oligosaccharide preferably means a sugar polymer containing at least two monosaccharide units (vide supra). The oligosaccharide can have a linear or branched structure containing monosaccharide units that are linked to each other by interglycosidic linkages.

The first aspect of the invention relates to a method of making an a-amino-aldehyde, comprising the step of reacting an a-hydroxy-aldehyde with at least two equivalents of an amine reagent in the presence of an acid catalyst. In the reaction the a-hydroxy group of the starting material is replaced by an amino group from the amine reagent.

The amine reagent is preferably of the formula HN-R 1 R 2 , wherein Ri and R 2 are H, alkyl, homoaryl, heteroaryl or benzyl, or i and 2 together form a divalent -(CI¾)n- moiety wherein n is 4-8 and a methylene group, preferably one methylene group, can be replaced by an oxygen, a sulphur or a > -R 3 moiety wherein R 3 is H or alkyl. Preferably, the amine reagent is a primary or a secondary amine. In a primary amine, Ri is H and R 2 is an optionally substituted alkyl, homoaryl, heteroaryl or benzyl, more preferably a methyl, ethyl, propyl, butyl, phenyl or benzyl. The most important primary amine is an optionally substituted benzyl amine wherein the optional substituents can be selected from the substituents listed at the "group removable by hydrogenolysis" of this invention (i.e., the most important primary amine is a primary amine of a group removable by hydrogenolysis, such as benzyl amine). A secondary amine (wherein Ri and R 2 are not H) can be a non-symmetrical (Ri≠ R 2 ) amine, a symmetrical (Ri = R 2 ) non-cyclic amine such as diethyl amine, dipropyl amine or dibenzyl amine or a symmetrical cyclic amine such as pyrrolidine, piperidine, piperazine, N-methyl piperazine, morpholine or thiomorpholine. Preferred amines are those having Ri and/or R 2 group(s) that can be removable by hydrogenolysis.

The a-hydroxy-aldehyde preferably is a sugar molecule from the group of mono-, di- or oligosaccharide aldoses. The monosaccharide aldose or the reducing monosaccharide moiety of the di- or oligosaccharide aldose is preferably a 5-6 carbon atom containing sugar (that is an aldopentose or an aldohexose) which is, in fact, known to occur in an open chain form or in a ring form (when the aldehyde group of the sugar makes a hemiacetal with one of its hydroxyl groups mostly in the form of a 5-6 membered ring). Formula 1 below represents a possible conformer of an α-hydroxy-aldehyde sugar:

wherein each R4, independently, is H or -OR6 wherein ¾ is H or a sugar

component, and is H, -COOH or -CH 2 R 4 .

In preferred monosaccharides of formula 1 all the R6-groups in R4 and R5 are H, and optionally R5 is H or -COOH. In preferred disaccharides of formula 1, one of the R6 -groups in R4 and 5 is a sugar component and the others are H. Among monosaccharides especially preferred are those in which R4 is -OR6, R5 is CH2-OR6 and Re is H, particularly D-glucose, D-galactose and D- mannose and most preferably D-glucose. Among disaccharides especially preferred are those in which one of the Re -groups in R4 and R5 is a sugar component selected from a pentose such as arabinose, ribose, xylose or lyxose and a hexose such as glucose, galactose, mannose, fucose, rhamnose, which sugar component is attached to the oxygen atom of the Regroup via an interglycosidic linkage, and the other Re-groups are H. A particularly preferred sugar component is galactose, especially when coupled to the 4-OH group of glucose, thereby forming lactose.

In carrying out the method of the first aspect of the invention, to initiate the conversion of an a-hydroxy-aldehyde to an a-amino-aldehyde an acidic catalyst is needed. The catalyst can be an inorganic protic acid such as HC1, HBr, sulphuric acid or phosphoric acid, or an organic protic acid such as formic acid, acetic acid, oxalic acid, or an optionally substituted methanesulphonic acid derivative (methanesulphonic acid, trifluoromethanesulphonic acid, etc.), or an optionally substituted benzenesulphonic acid derivative (benzenesulphonic acid, p-toluenesulphonic acid, etc.), or a polymer bound sulphonic acid (i.e., an ion exchange resin), or a Lewis-acid such as AICI 3 , ZnC¾, CuBr 2 , MgC¾ or BF 3 -etherate. Preferably, the catalyst is an acid addition salt of the amine reagent selected from a halide, hydrogen phosphate, N-benzyl-carbamate, bicarbonate or carbonate salt, or carbon dioxide adduct of the amine reagent, particularly a hydrochloride.

The a-amino-aldehyde product of the method of the first aspect is preferably a sugar molecule selected from the group of mono-, di- and oligosaccharide 2-amino-2-deoxy-aldoses and their N-glycosides. The monosaccharide 2-amino-2-deoxy-aldose or the reducing monosaccharide 2-amino-2-deoxy-aldose moiety of the di- or oligosaccharide aldose can be a 5-6 carbon atom containing 2-amino-2-deoxy sugar derivative which are, in fact, known to occur in an open chain form or in a ring form (when the aldehyde group of the sugar makes a hemiacetal with one of its hydroxyl groups mostly in the form of a 5-6 membered ring). Formula 2a below represents a possible conformer of a 2-amino-2-deoxy-aldose, and formula 2b depicts its N-glycoside:

wherein Ri, R2, R3, R* and R5 are defined as above.

Preferably, a compound of formula 2a or 2b is a monosaccharide in which all the R6-groups in R4 and R5 are H, and optionally R5 is H or -COOH, or a disaccharide wherein one of the Regroups in R4 and R5 is a sugar component and the others are H. In the series of monosaccharides those in which R4 is -OR 5 , 5 is CH 2 -OR 6 and 6 is H are more preferred, among which compounds 2-amino-2-deoxy-D-glucose derivatives, 2-amino-2-deoxy-D-galactose derivatives and 2-amino-2-deoxy-D-mannose derivatives, particularly 2-amino-2-deoxy-D-glucose derivatives are even more preferred. In the series of disaccharides, one of the Re-groups in R4 and R5 is a sugar component selected from a pentose such as arabinose, ribose, xylose or lyxose and a hexose such as glucose, galactose, mannose, fucose, rhamnose, which sugar component is attached to the oxygen atom of the Regroup via an interglycosidic linkage. More preferably, the sugar component is galactose which is, in a more preferred embodiment, coupled to the 4-OH group of 2-amino-2- deoxy-glucose derivative forming thus a 2-amino-2-deoxy-lactose (lactosamine) derivative.

In carrying out the method of the first aspect of the invention, at least two molar equivalents of the amine reagent are needed. However, for better productivity and conversion, it is

advantageous to use more than two equivalents of the amine reagent, that is, at least three equivalents, preferably at least four equivalents, especially at least five equivalents of the amine reagent. As to the amount of the acidic catalyst, it should preferably be used in a quantity such that at least two molar equivalents of the amine reagent remain after its partial neutralization by the acidic catalyst. It is believed that the chemical process of this method takes place via at least the following two steps: i) an Amadori rearrangement of the starting a-hydroxy-aldehyde with one mole of the amine reagent resulting in an intermediary a-amino-ketone, and ii) a Heyns-type rearrangement of the a-amino-ketone as a result of its reaction with another mole of the amine reagent to form an a- amino-aldehyde. This finding is surprising as there is no example, suggestion or hint in the prior art of which the inventors are aware that an α-hydroxy-aldehyde, particularly an aldose, can be converted directly to an a-amino-aldehyde, particularly a 2-amino-2-deoxy-aldose derivative, or an N-glycoside thereof. This process can be depicted by Scheme 1 below, wherein Ri, R 2 , R3, R4 and R5 are defined as above.

Scheme 1.

Although the intermediate of formula 3 is drawn in open-chain form in Scheme 1, it is believed more likely that this intermediate occurs in cyclic hemiketal form. With regard to the original configuration of C-2 of the starting α-hydroxy-aldehyde of formula 1 in Scheme 1, it is lost as this starting material is converted into a compound having prochiral carbon atom at C-2 (i.e. the keto-group in intermediate of formula 3), which carbon atom becomes chiral again in the second part of the reaction, thus the formation of both 2-epimers of the product a-amino-aldehyde can be expected

In line with this postulated stereochemical outcome of the reaction, when the starting a- amino-aldehyde is a chiral molecule with respect to the a-carbon atom, an a-epimeric pair of the product α-amino-aldehyde can be obtained, in which the proportion of the epimers depends on the reaction condition and presumably the thermodynamically most stable stereoisomer forms either exclusively or predominantly. For example, from optionally glycosylated D-glucose, both 2-amino- 2-deoxy-D-glucose and -D-mannose derivatives, optionally in glycosylated form, can be obtained, but mainly the D-gluco compounds (see Scheme 2).

H +

Scheme 2.

As shown in Scheme 2, mostly the formation of the 2-amino-2-deoxy-aldose derivative N- glycoside (compounds of formula 2b) can be expected. Preferably, the reaction is driven in the presence of at least three equivalents of the amine reagent, more preferably of at least four equivalents, even of at least five equivalents of the amine reagent so that an N-glycoside of formula 2b forms as a primary product. Compounds of formula 2a can also be found in the reaction mixture, especially when some water is present or the work-up procedure comprises aqueous treatment (e.g. extraction). Compounds of formula 2b (also named: l ,2-diamino-l,2-dideoxy-aldose derivatives) have the advantageous feature that they can be isolated more easily, even in crystalline form, from the reaction mixture than the 2-amino-2-deoxy-aldose derivatives of formula 2a (see WO

2009/059945). The separated compounds of formula 2b can then be hydrolysed under well-known conditions (i.e. acidic hydrolysis) to compounds of formula 2a, namely the anomeric -NR1R2 group is acid labile. In this reaction water - which is present in the reaction milieu as reagent - can serve as solvent or co-solvent as well. Organic protic or aprotic solvents which are stable under acidic conditions and miscible fully or partially with water such as Ci-Ce alcohols, acetone, THF, dioxane, ethyl acetate, MeCN, etc. can be used in a mixture with water. The acids used are generally inorganic protic acids selected from but not limited to acetic acid, trifluoroacetic acid, HC1, formic acid, sulphuric acid, perchloric acid, oxalic acid, -toluenesulphonic acid,

benzenesulphonic acid and cation exchange resins, and organic acids including but not limited to acetic acid, formic acid, chloroacetic acid and oxalic acid, which can be present in from catalytic amount to large excess. The hydrolysis can be conducted at temperatures between 20 °C and reflux until reaching completion which takes from about 2 hours to 3 days depending on temperature, concentration and pH. Preferably, the hydrolysis is performed in an alcohol, more preferably in methanol, ethanol or isobutanol, by addition of concentrated or diluted aqueous HCl-solution, and the pH is kept at around 2-4. Under such conditions the hydrolysis is typically complete within 2-3 hours at room temperature.

If it is desired to produce a salt of one of the compounds of formula 2a or 2b by the method of the first aspect of the invention, the free base can be converted into its acid addition salt in a conventional manner, using inorganic or organic acids or salts. Solvents such as acetone, water, dioxane, DMSO, THF, DMF, alcohols, MeCN, and mixtures thereof and inorganic acids such as HC1, H2SO4, HNO3 and H3PO4, in concentrated form or diluted in water or other solvents, such as methanol, ethanol or dioxane, can be used. Organic acids such as formic acid, acetic acid, and oxalic acid can also be used. The salts of these acids with a base whose basicity is weaker than that of a compound of formula 2a or 2b can be used as well. Products are typically obtained by selective precipitation by adding apolar solvents such as diethyl ether, diisopropyl ether, acetone or alcohols, or by crystallization without any chromatography.

In a preferred embodiment of this method, lactose is reacted with optionally substituted benzyl amine, preferably benzyl amine, which is used in at least 3-4-fold molar excess. Preferably, lactose is added to the amine reagent acting also as the solvent, followed by the addition of the salt of the amine reagent at about 20-30 °C. Alternatively, lactose is added to the mixture of the amine reagent and its salt. The salt is usually the hydrochloride of the amine reagent, preferably benzylammonium hydrochloride, and can be used from catalytic amount to one molar equivalent (in terms of the starting sugar) The reaction mixture is preferred to be heated to around 55-85 °C, preferably around 70 °C, for some hours to 2 days. The main product is l,2-dideoxy-l,2- dibenzylamino-lactose optionally accompanied by traces of l,2-dideoxy-l,2-dibenzylamino- epilactose, which can be transformed into 2-benzylamino-2-deoxy-lactose (optionally with 2- benzylamino-2-deoxy-epilactose) by acidic hydrolysis. When D-glucose is used, l,2-dideoxy-l,2- dibenzylamino-D-glucose (optionally with l,2-dideoxy-l,2-dibenzylamino-D-mannose) and 2- deoxy-2-benzylamino-D-glucose (optionally with 2-dideoxy-2-benzylamino-D-mannose) can be produced. The use of optionally substituted benzyl amines (HN-R R 2 wherein Ri is H and R 2 is a group removable by hydrogenolysis) is particularly advantageous in that the R 2 -group in a compound of formula 2a or both Ri-groups in a compound of formula 2b can be easily removed by hydrogenolysis to give the free aminodeoxy sugar This reaction can be suitably carried out in a protic solvent or in a mixture of protic solvents. The protic solvent can be water, acetic acid or a Ci- Ce alcohol. A mixture of one or more protic solvents with one or more suitable aprotic organic solvents partially or fully miscible with the protic solvent(s) (such as THF, dioxane, ethyl acetate or acetone) can also be used. Water, one or more C \ -Ce alcohols, or a mixture of water and one or more C1-C6 alcohols are preferably used as the solvent system. Solutions containing the carbohydrate derivatives in any concentration or suspensions of the carbohydrate derivatives in the solvent(s) used are also applicable. The reaction mixture is stirred at a temperature in the range of 10-100 °C, preferably between 20-50 °C, in a hydrogen atmosphere of 1-50 bar absolute (100 to 5000 kPa) in the presence of a catalyst such as palladium, Raney nickel or any other appropriate metal catalyst, preferably palladium on charcoal or palladium black, until reaching the completion of the reaction. Transfer hydrogenation can also be performed, when the hydrogen is generated in situ from cyclohexene, cyclohexadiene, formic acid or ammonium formate. Addition of organic or inorganic bases or acids and/or basic and/or acidic ion exchange resins can also be used to improve the kinetics of the hydrogenolysis. The use of basic substances is especially preferred when starting from a compound where halogen substituent(s) is/are present on the substituted benzyl

moiety/moieties and/or the formation of deoxy-amino sugar base is desirable. Preferred organic bases include, but are not limited to, triethylamine, diisopropyl ethylamine, ammonia, ammonium carbamate and diethylamine. An organic or an inorganic acid is preferably used as a co-solvent or additive in cases when salts of the deoxy-amino sugar are the intended products. Preferred acids include, but are not limited to, formic acid, acetic acid, propionic acid, chloroacetic acid, dichloroacetic acid, trifluoroacetic acid, HC1 and HBr. When starting from a compound of formula 2b, an acid can be used, while the hydrolysis and the hydrogenolysis simultaneously take place. The conditions proposed above allow simple, convenient and delicate production of a pure amino sugar or its salt. Preferably, conducting the method of the first aspect of the invention starting from D- glucose followed by catalytic hydrogenolysis gives 2-amino-2-deoxy-D-glucose and/or 2-amino-2- deoxy-D-mannose or their salts which can serve as precursors for making biologically important amino sugar derivatives such as N-acetyl-D-glucosamine, N-acetyl-D-mannosamine, sialic acid, etc. (see e.g. WO 2012/140576). Likewise, if lactose is the starting material in this method followed by catalytic hydrogenolysis, lactosamine or its salts can be obtained which is a precursor of N- acetyl-lactosamine (see e.g. WO 2009/059945).

The second aspect of the invention relates to a method for making an a-amino-aldehyde, comprising the steps of:

a) reacting an a-hydroxy-aldehyde with an amine reagent to obtain an a-amino-ketone, and b) reacting the a-amino-ketone obtained in step a) with an amine reagent, preferably with the same amine reagent employed in step a), to obtain an a-amino-aldehyde.

The a-hydroxy-aldehyde used as starting material in step a) preferably is a sugar molecule from the group of mono-, di- or oligosaccharide aldoses. The monosaccharide aldose or the reducing monosaccharide moiety of the di- or oligosaccharide aldose is preferably a 5-6 carbon atom containing sugar (that is an aldopentose or an aldohexose) which is, in fact, known to exist in an open chain form or in a ring form (when the aldehyde group of the sugar makes a hemiacetal with one of its hydroxyl groups mostly in the form of a 5-6 membered ring). Formula 1 below represents a possible conformer of an α-hydroxy-aldehyde sugar:

wherein each R4, independently, is H or -OR^ wherein R^ is H or a sugar

component, and R5 is H, -COOH or -CH 2 R 4 .

In preferred monosaccharides of formula 1 all the R^-groups in R4 and R5 are H, and optionally R5 is H or -COOH. In preferred disaccharides of formula 1, one of the Re-groups in R4 and R5 is a sugar component and the others are H. Among monosaccharides especially preferred are those in which R4 is -OR 6 , R5 is CH 2 -OR 6 and ¾ is H, particularly D-glucose, D-galactose and D- mannose and most preferably D-glucose. Among disaccharides especially preferred are those in which one of the Re-groups in R4 and R5 is a sugar component selected from a pentose such as arabinose, ribose, xylose or lyxose and a hexose such as glucose, galactose, mannose, fucose, rhamnose, which sugar component is attached to the oxygen atom of the Regroup via an interglycosidic linkage, and the other R 6 -groups are H. A particularly preferred sugar component is galactose, especially when coupled to the 4-OH group of glucose, thereby forming lactose.

In carrying out step a) of the method of the second aspect of this invention, the same amine reagent, preferably of the formula HN-R 1 R 2 , wherein i and R 2 are as defined above, can be used as in the method of the first aspect of the invention. Preferably, the amine reagent is a primary or a secondary amine. In a primary amine, Ri is H and R 2 is optionally substituted alkyl, homoaryl, heteroaryl or benzyl, more preferably methyl, ethyl, propyl, butyl, phenyl or benzyl. The most prominent primary amine is optionally substituted benzyl amine wherein the optional substituents can be selected from those listed at "group removable by hydrogenolysis" above (that is, the most prominent primary amine is a primary amine of a group removable by hydrogenolysis, such as benzylamine). A secondary amine (Ri and R 2 are not H) can be a non-symmetrical (Ri≠ Ri) amine, a symmetrical (Ri = R2) non-cyclic amine such as diethyl amine, dipropyl amine, dibenzyl amine etc. or a symmetrical cyclic amine such as pyrrolidine, piperidine, piperazine, N-methyl piperazine, morpholine, thiomorpholine, etc. So the most preferred amines to be used in step a) of the method of the second aspect are those having i and/or R 2 group(s) that can be removable by

hydrogenolysis.

The a-amino-ketone formed in step a) of the method according to the second aspect is preferably a sugar molecule selected from the group of mono-, di- and oligosaccharide 1 -amino- 1- deoxy-ketoses. The monosaccharide 1-amino-l-deoxy-ketose or the reducing monosaccharide 1- amino- 1-deoxy-ketose moiety of the di- or oligosaccharide ketose can be a 5-6 carbon atom containing 1-amino-l-deoxy sugar derivative which is, in fact, known to exist in an open chain form or in a ring form (when the keto group of the sugar makes a hemiketal with one of its hydroxyl groups mostly in the form of a 5-6 membered ring). This type of compound is usually called an Amadori product (and the reaction is an Amadori rearrangement); formula 3 in Scheme 3 below represents the open chain form of a 1-amino-l-deoxy-ketose (Ri, R 2 , R3, 4 and R5 are defined as above), and as examples, the Amadori products of D-glucose and lactose are also shown:

of a Ό-fructo derivative (as compound 3)

of a lactulose derivative (as compound 3)

Scheme 3.

In a preferred embodiment, the compound of formula 3 is a monosaccharide (all the Regroups in R4 and R5 are H, and optionally R5 is H or -COOH) or a disaccharide (one of the R - groups in R4 and R5 is a sugar component and the others are H). In the series of monosaccharides, those in which R4 is -OR6, R5 is CH 2 -OR6 and R¾ is H are more preferred, among which compounds 1 -amino- l-deoxy-fructose derivatives are more preferred. In the series of disaccharides, it is preferred that one of the Regroups in R4 and R5 is a sugar component selected from a pentose such as arabinose, ribose, xylose or lyxose and a hexose such as glucose, galactose, mannose, fucose, rhamnose, which sugar component is attached to the oxygen atom of the Regroup via an interglycosidic linkage. More preferably, the sugar component is galactose which is, in a more preferred embodiment, coupled to the 4-OH group of 1 -amino- l-deoxy-fructose derivative forming thus a 1 -amino- 1-deoxy-lactulose derivative. To initiate the conversion of an a-hydroxy-aldehyde to an a-amino-ketone in step a) of the method of the second aspect a catalyst is needed. The catalyst can be an organic or inorganic protic acid as described above or an acid addition salt of ammonia or of the amine reagent selected from a halide, hydrogen phosphate, N-benzyl-carbamate, bicarbonate and carbonate salt, preferably ammonium chloride or the chloride salt of the amine reagent, or an active methylene compound such as malonate ester (e.g. diethyl malonate), phenyl acetone, acetyl acetone, or diphenyl methane.

In carrying out step a) of the method of the second aspect, one can proceed in two ways.

According to a first way, step a) involves heating the starting aldose, the amine reagent and the catalyst together. The amine reagent is taken in equimolar amount or only in a slight excess (between 1-1.5 equivalents to the aldose) and can be used as solvent (if it is a liquid). The reaction mixture can be diluted with a polar solvent such as water, alcohol (methanol, ethanol, propanol, isopropanol, isobutanol) or dioxane. The catalyst is preferably selected from an inorganic protic acid, usually HCl solution, and an organic acid, usually acetic acid or oxalic acid, particularly if the amine reagent is an aromatic or benzyl amine, and is preferably used in an amount that brings the pH of the reaction mixture under 7. Alternatively, the catalyst is preferably an active methylene compound which can be used in excess when the amine reagent is an aliphatic primary amine or a secondary amine. The reaction mixture is preferably heated to around 55-85 °C, more preferably around 60-70 °C for 1-24 hours.

According to a second way, step a) involves heating the aldose and the amine reagent together, in relative amounts as disclosed above for the first way, and optionally dissolved in alcohol, preferably at around 55-85 °C, more preferably around 55-60 °C, preferably for 1-4 hours, in the absence of a catalyst. As a result, a glycosyl amine of formula 4, wherein R 1; R 2 , R3, R4 and R5 are as defined above, is formed, which can then be converted into the Amadori product of formula 3 by addition of the catalyst and upon further heating, preferably at around 50-70 °C, preferably for 1-18 hours. Alternatively, in order to achieve a better yield, the starting aldose is treated with a high excess of amine in the absence of catalyst, the so-formed glycosyl amine of formula 4 is separated by precipitation or crystallization, the excess of the reagent amine is washed away, and the glycosyl amine is then dissolved in alcohol or dioxane and treated with an inorganic or organic acid or active methylene compound, preferably when heated as above, to initiate its rearrangement to a compound of formula 3. If the glycosyl amine of formula 4 cannot be, or is not intended to be, separated before inducing the Amadori rearrangement, at least as much acid catalyst, preferably an inorganic or organic acid, is added to neutralize the excess of amine and make the pH acidic. This two-step procedure to make a 1 -amino- 1 -deoxy-ketose of formula 3 from an aldose via the glycosyl amine is shown in Scheme 4 below.

Scheme 4.

In step b) of the method of the second aspect, the a-amino-ketone obtained, e.g. a compound of formula 3, is converted into an a-amino-aldehyde and/or, if the resulting a-amino-aldehyde is a sugar derivative able to form a cyclic hemiacetal, an N-glycoside thereof. This is done practically under the conditions characteristic for the Heyns rearrangement. Surprisingly, a 1 -amino- 1 -deoxy- ketose derivative (Amadori product) of formula 3 can undergo Heyns rearrangement to give a 2- amino-2-deoxy-aldose derivative such as an α-amino-aldehyde and/or N-glycoside thereof. The a- amino-aldehyde as the product of the Heyns rearrangement is preferably a sugar molecule selected from the group of mono-, di- and oligosaccharide 2-amino-2-deoxy-aldoses and their N-glycosides. The monosaccharide 2-amino-2-deoxy-aldose, or the reducing monosaccharide 2-amino-2-deoxy- aldose moiety of the di- or oligosaccharide aldose can be a 5-6 carbon atom containing 2-amino-2- deoxy sugar derivative which are, in fact, known to exist in an open chain form or in a ring form (when the aldehyde group of the sugar makes a hemiacetal with one of its hydroxyl groups, in general in the form of a 5 or 6 membered ring). Formula 2a below represents one of the possible conformers of a 2-amino-2-deoxy-aldose, and formula 2b depicts its N-glycoside:

wherein Ri, R 2 , R3, R4 and R5 are defined as above.

Preferably, the compound of formula 2a or 2b is a monosaccharide (all the Regroups in R4 and 5 are H, and optionally R5 is H or -COOH) or a disaccharide (one of the Rs-groups in R4 and R5 is a sugar component and the others are H). In the series of monosaccharides in which R4 is -ORe, 5 is CH 2 -OR 6 and Re is H are more preferred, among which compounds 2-amino-2-deoxy- D-glucose derivatives, 2-amino-2-deoxy-D-galactose derivatives and 2-amino-2-deoxy-D-mannose derivatives, particularly 2-amino-2-deoxy-D-glucose derivatives are even more preferred. In the series of di saccharides, one of the Re-groups in R 4 and R5 is a sugar component selected from a pentose such as arabinose, ribose, xylose or lyxose and a hexose such as glucose, galactose, mannose, fucose, rhamnose, which sugar component is attached to the oxygen atom of the Regroup via an interglycosidic linkage. More preferably, the sugar component is galactose which is, in a more preferred embodiment, coupled to the 4-OH group of 2-amino-2-deoxy-glucose derivative forming thus a 2-amino-2-deoxy-lactose (lactosamine) derivative.

Step b) of the method of the second aspect can be practically carried out in three ways.

According to a first way, step b) involves the sub-steps of: I) treating a compound of formula 3 obtained in step a) above with an amine reagent, preferably of formula HN-R 1 R 2, to yield a ketosyl amine derivative of formula 5 (wherein Ri, R 2 , R3, R 4 and R5 are defined as above), II) isolating the ketosyl amine derivative of formula 5, and III) treating the isolated ketosyl amine derivative of formula 5 with an acid to obtain a compound of formula 2a and/or 2b. In step I), a compound of formula 3 obtained in step a) is reacted with at least an equimolar amount, and preferably an excess (ca. 3-10 equiv.), of an amine reagent of formula HN-R 1 R 2 wherein Ri and R 2 are as defined above at step a). The amine reagent can be the same as, or different from, that used in step a), but preferably the amine reagent is identical to that used in step a), and is more preferably an optionally substituted benzyl amine, particularly benzyl amine. The amine reagent - if liquid - can also serve as a solvent, or a concentrated solution of the amine reagent in alcohol, dioxane, THF, DMF, or another suitable solvent can be used. Preferably, a compound of formula 3 obtained in step a) is added to the amine reagent as solvent at about 0 °C, and then the mixture is allowed to warm to room temperature or slowly heated up to 40 °C, so that the starting material is consumed. The reaction is continued until consumption of the starting material is observed by TLC, which is typically observed within 24 h, usually within 18-20 hours. In step II), the remaining amine reagent or the excess of the amine reagent is removed from the crude ketosyl amine derivative of formula 5 before adding acid to initiate the rearrangement reaction III). Apolar solvents not dissolving the intermediate glycosyl amine derivative, mainly lower hydrocarbons such as pentanes, hexanes, or heptanes, or mixtures thereof such as petroleum ether, are suitable to extract the amine reagent. As the ketosyl amine derivative of formula 5 formed in the reaction is less soluble in apolar solvents, the organic layer containing the amine reagent can be easily separated. Hence, any excess of the amine reagent is preferably washed away by using petroleum ether in step II). Preferably, the suspension/emulsion formed after addition of the apolar solvent is frozen at a temperature of between -20 and -25 °C and the supernatant organic phase is decanted. The supernatant organic phase is found not to contain any significant quantity of carbohydrate-like compound. The washing procedure can be repeated several times. The glycosyl amine derivative can be precipitated and/or crystallized but can also be used directly in step III). In step III), the crude/crystalline glycosyl amine derivative is dissolved in alcohol, dioxane, THF, DMF or a mixture thereof, preferably in alcohol, more preferably in methanol, and an acid is added to promote rearrangement. Preferably, in step III) the crude/crystalline ketosyl amine derivative of formula 5 is taken up in alcohol, such as methanol, isopropanol or isobutanol, followed by addition of an acid. The acid can be used in any amount from a catalytic amount to a large excess. The acid can be an inorganic protic acid such as HC1, HBr, sulphuric acid or phosphoric acid, or an organic protic acid such as formic acid, acetic acid, oxalic acid, an optionally substituted methanesulphonic acid derivative, an optionally substituted benzenesulphonic acid derivative, a polymer bound sulphonic acid (i.e., an ion exchange resin), or a Lewis-acid such as AICI 3 , ZnC¾, CuBr 2 , MgCl 2 or BF 3 -etherate, preferably glacial acetic acid. The reaction typically takes place at room temperature and is completed within several hours, such as up to 8 hours, and preferably within 2-4 hours, and a compound of formula 2a and/or 2b is produced. Scheme 5 shows the process of step b) where the ketosyl amine derivative of formula 5 is depicted in furanosyl form, but it can occur in pyranosyl form as well if R5 contains a free OH-group. When the acid catalyst in step III) is an organic or inorganic protic acid in aqueous solution the formation of a compound of formula 2a is preferable, otherwise under non-aqueous conditions the production of compound of formula 2b can be expected as the exclusive or the major product. It should also be noted that any compound of formula 2a and 2b - with regard to the C-2 carbon atom - can be a single stereoisomer or a mixture of 2-epimers.

Scheme 5. According to a second way, step b) involves the sub-step of treating a compound of formula 3, obtained in step a) above, with an amine reagent as described above in sub-step I), preferably in around 7-8 molar excess. This sub-step involves heating at 30-50 °C for 6-24 hours while the water formed during the conversion of compound of formula 3 to that of formula 5 is removed by azeotropic distillation. Then, the crude ketosyl amine derivative of formula 5 is taken up in an alcohol, such as methanol, isopropanol or isobutanol, followed by adding 0.2-1.0 equivalents of an acid addition salt of the amine reagent, preferably its halide. The resulting mixture is heated to 30- 40 °C and stirred for at least a day to give a compound of formula 2b either as a single stereoisomer or as a pair of C-2 epimers.

According to the third way, step b) involves the sub-step of treating a compound of formula

3 with an amine reagent and a salt thereof. In this sub-step, a compound of formula 3 is reacted with an excess, preferably 2-10 equiv., more preferably 3-4 equiv., of the amine reagent in the presence of a salt of the amine reagent. The amine reagent can be the same or different from that used in step a) but preferably is identical to that used in step a), and more preferably is an optionally substituted benzyl amine, particularly benzyl amine. The amine reagent - if liquid - can also serve as a solvent, or a concentrated solution of the amine reagent in alcohol, dioxane, THF, DMF, or another suitable solvent can be used. The salt of the amine reagent preferably refers to a halide, hydrogen phosphate, N-benzyl-carbamate, bicarbonate or carbonate salt, or carbon dioxide adduct of the amine reagent, more preferably the salt of an optionally substituted benzyl amine, particularly benzyl ammonium hydrochloride, and it is used in 0.2-1.0 equivalents. Preferably, a compound of formula 3 is added to the amine reagent acting also as the solvent, followed by the addition of the salt of the amine reagent at about 20-30 °C. Alternatively, a compound of formula 3 is added to the mixture of the amine reagent and its salt. The reaction is continued until consumption of the starting material as monitored by TLC and product equilibrium is reached as monitored by HPLC, which is typically observed within 6-8 days, and a compound of formula 2b forms either as a single stereoisomer or as a pair of C-2 epimers.

Compounds of formula 2b have the advantageous feature that they can be isolated more easily, in some cases in crystalline form, from the reaction mixture than the 2-amino-2-deoxy- aldose derivatives of formula 2a (see WO 2009/059945). The separated compounds of formula 2b can then be hydrolysed under well-known conditions (i.e. acidic hydrolysis) to compounds of formula 2a, namely the anomeric -NR 1 R 2 group is acid labile. In this reaction water - which is present in the reaction milieu as reagent - can serve as solvent or co-solvent as well. Organic protic or aprotic solvents which are stable under acidic conditions and miscible fully or partially with water such as Ci-Ce alcohols, acetone, THF, dioxane, ethyl acetate, MeCN, etc. can be used in a mixture with water. The acids used are generally inorganic protic acids selected from but not limited to acetic acid, trifluoroacetic acid, HC1, formic acid, sulphuric acid, perchloric acid, oxalic acid, /?-toluenesulphonic acid, benzenesulphonic acid and cation exchange resins, and organic acids including but not limited to acetic acid, formic acid, chloroacetic acid and oxalic acid, which can be present in from catalytic amount to large excess The hydrolysis can be conducted at temperatures between 0 °C and reflux, preferably between 20 °C and reflux, until reaching completion which takes from about 2 hours to 3 days depending on temperature, concentration and pH. Preferably, the hydrolysis is performed in an alcohol, more preferably in methanol, ethanol or isobutanol, by addition of concentrated or diluted aqueous HCl-solution, and the pH is kept at around 2-4. Under such conditions the hydrolysis is typically complete within 2-3 hours at room temperature.

Where it is desired to produce a salt of one of the derivatives of formula 2a or 2b, the derivative can be converted into its acid addition salt in a conventional manner, using inorganic or organic acids or salts. Solvents such as acetone, water, dioxane, DMSO, THF, DMF, alcohols, MeCN, and mixtures thereof and inorganic acids such as HC1, H 2 SO 4 , HNO 3 and H 3 PO 4 , in concentrated form or diluted in water or other solvents, such as methanol, ethanol or dioxane, can be used. Organic acids such as formic acid, acetic acid, and oxalic acid can also be used. The salts of these acids with a base whose basicity is weaker than that of a derivative of formula 2a or 2b can be used as well. Products are typically obtained by selective precipitation by adding apolar solvents such as diethyl ether, diisopropyl ether, acetone or alcohols, or by crystallization in high yield without any chromatography

In a preferred realization of the method according to the second aspect, in step a) D-glucose or lactose is reacted with an optionally substituted benzyl amine reagent, preferably benzyl amine. Acetic acid promoting the rearrangement can be added at the beginning of the reaction or after the formation of the respective glycosyl amine, and 1-benzylamino-l-deoxy-D-fructose or 1- benzylamino-l-deoxy-lactulose, respectively, is obtained. If the acid is used in more than 1 equivalent, the corresponding salt of the Amadori product can be obtained. In step b), the so-formed Amadori product, either the free base or the salt, is then reacted with the optionally substituted benzyl amine, preferably benzyl amine, which is used in 3-8-fold molar excess at about 20-40 °C for 2-6 days. The reaction is promoted by a salt of the amine reagent, usually the hydrochloride of the amine reagent, preferably benzylammonium hydrochloride, in from catalytic amount to one molar equivalent (in terms of the Amadori product). The main product is l,2-dideoxy-l,2- dibenzylamino-D-glucose (optionally with l ,2-dideoxy-l ,2-dibenzylamino-D-mannose) or 1 ,2- dideoxy-l,2-dibenzylamino-lactose (optionally accompanied by traces of l ,2-dideoxy- l ,2- dibenzylamino-epilactose), respectively, which can be transformed into 2-deoxy-2-benzylamino-D- glucose (optionally with 2-dideoxy-2-benzylamino-D-mannose) or 2-benzylamino-2-deoxy-lactose (optionally with 2-benzylamino-2-deoxy-epilactose), respectively, by acidic hydrolysis.

The use of optionally substituted benzyl amine reagents (HN-R 1 R 2 wherein ¾ is H and R 2 is a group removable by hydrogenolysis) is particularly advantageous in the regard that the R 2 -group in a compound of formula 2a or both R 2 -groups in a compound of formula 2b can be removed smoothly by hydrogenolysis to give the free 2-amino-2-deoxy sugars. This reaction typically takes place in a protic solvent or in a mixture of protic solvents. The protic solvent can be water, acetic acid or a Ci-Ce alcohol. A mixture of one or more protic solvents with one or more suitable aprotic organic solvents partially or fully miscible with the protic solvent(s) (such as THF, dioxane, ethyl acetate or acetone) can also be used. Water, one or more C1-C6 alcohols, or a mixture of water and one or more C1-C6 alcohols are preferably used as the solvent system. Solutions containing the carbohydrate derivatives in any concentration or suspensions of the carbohydrate derivatives in the solvent(s) used are also applicable. The reaction mixture is stirred at a temperature in the range of 10- 100 °C, preferably between 20-50 °C, in a hydrogen atmosphere of 1-50 bar absolute (100 to 5000 kPa) in the presence of a catalyst such as palladium, Raney nickel or any other appropriate metal catalyst, preferably palladium on charcoal or palladium black, until reaching the completion of the reaction. Transfer hydrogenation can also be performed, when the hydrogen is generated in situ from cyclohexene, cyclohexadiene, formic acid or ammonium formate. Addition of organic or inorganic bases or acids and/or basic and/or acidic ion exchange resins can also be used to improve the kinetics of the hydrogenolysis. The use of basic substances is especially preferred when starting from a compound where halogen substituent(s) is/are present on the substituted benzyl

moiety/moieties and/or the formation of deoxy-amino sugar base is desirable. Preferred organic bases include, but are not limited to, triethylamine, diisopropyl ethylamine, ammonia, ammonium carbamate and diethylamine. An organic or an inorganic acid is favourably used as a co-solvent or additive in cases when salts of the deoxy-amino sugar are the intended products. Preferred acids include, but are not limited to, formic acid, acetic acid, propionic acid, chloroacetic acid, dichloroacetic acid, trifluoroacetic acid, HC1 and HBr. When starting from a compound of formula 2b, an acid can be used, while the hydrolysis and the hydrogenolysis simultaneously take place. The conditions proposed above allow simple, convenient and delicate removal of the solvent(s) giving rise to pure amino sugar or its salt. In a preferred realization, conducting the method of the first or second aspect of the invention disclosed above from D-glucose followed by catalytic

hydrogenolysis of the Heyns rearranged product gives 2-amino-2-deoxy-D-glucose and/or 2-amino- 2-deoxy-D-mannose or their salts which can serve as precursor for making biologically important amino sugar derivatives such as N-acetyl-D-glucosamine, N-acetyl-D-mannosamine, sialic acid, etc (see e g. WO 2012/140576). Likewise, if lactose is employed in the method of the first or second aspect of the invention disclosed above followed by catalytic hydrogenolysis of the Heyns rearranged product, lactosamine or its salts can be obtained which is a precursor of N-acetyl- lactosamine (see e.g. WO 2009/059945).

In accordance with a third aspect of this invention, derivatives of the following formula 6, which are preferred embodiments of compounds of formula 3 above, and their salts are provided

6

wherein Ri a and R. 2a are each H or a group removable by hydrogenolysis, each R4 a , independently, is H or -OR6 a ; Rs a is H, -COOH or -CH 2 R4 a ; and Re a is H or a sugar component, provided that at least one of R la or R 2a is a group removable by

hydrogenolysis and provided that there is one R6 a in Rt a and Rs a which is a sugar

component and any other Re a are H.

A compound of formula 6 and salts thereof can be characterized as crystalline solids, oils, syrups, precipitated amorphous material or spray dried products If crystalline, they can exist either in anhydrous or in hydrated crystalline forms by incorporating one or several molecules of water into their crystal structures. Similarly, compounds of formula 6 and salts thereof can exist as crystalline substances incorporating ligands such as organic molecules and/or ions into their crystal structures. The compounds of formula 6 and salts thereof includes anomeric mixtures of a- and β- anomers and/or the pure form of the a- or the β-anomers.

In certain embodiments, the third aspect of the invention excludes (4-fluorophenyl)-{ [2,3,5- trihydroxy-4-(3,4,5-trihydroxy-6-hydroxymethyl-tetrahydropyr an-2-yloxy)-tetrahydropyran-2- ylmethyl]amino}methane and/or salts thereof. In certain embodiments, the third aspect of the invention excludes compounds according to formula 6 wherein Ri a is H, R 2a is a benzyl or naphthylmethyl optionally substituted by 1-3 substituents selected from the group consisting of: amino, halogen, hydroxy, nitro, Ci-Ce alkylamino, Ci-Ce dialkylamino, C1-C6 alkyl and Ci-Ce alkoxy, Rs a is CH2OH and R4 a is OR6a where one R6 a is H and the other ¾ a is a monosaccharide.

It should be emphasized that compounds of formula 6 can exist, both in solution and as solids, in cyclic hemiacetal form, especially pyranose and furanose forms (both having a- and β-anomers), as well as in open chain form. The relative proportion of the forms depends on the nature of the solvent(s), concentration, temperature and/or condition(s) under which solidification, precipitation, crystallization or other means of solvent removal is carried out.

Preferably, the compounds of formula 6 are characterized by formula 6a

6a wherein Ri a , R2 a , R4 a and Rs a are as defined above.

More preferably, Ri a is H and R 2 a is an optionally substituted benzyl, more preferably benzyl. Even more preferably, there is one R^-group in R4 a and Rs a which is a sugar component selected from a pentose such as arabinose, ribose, xylose or lyxose and a hexose such as glucose, galactose, mannose, fucose, rhamnose, and which sugar component is attached to the oxygen atom of the Regroup via an interglycosidic linkage, and any other R^-groups are H. More preferably, the sugar component is galactose which is, in a more preferred embodiment, coupled to the 4-OH group of 1- benzylamino- l -deoxy- fructose derivative forming thus 1-benzylamino-l-deoxy-lactulose.

A fourth aspect of the invention relates to the use of the compounds of the third aspect of the invention as a synthetic intermediate, in particular as a synthetic intermediate in the formation of a- amino-aldehydes, such as in the method of the second aspect of the invention. EXAMPLES

Example 1

A suspension of lactose monohydrate (10 g) and benzylammonium hydrochloride (4.0 g) in benzylamine (20 ml) was stirred at 70 °C for 5 hours under reduced pressure (100-200 mbar). The obtained dark red-brown solution was partially evaporated at to give a thick oil (26 g). It was homogenized with 2-methyltetrahydrofuran (Me-THF, 30 ml) and allowed to stand in the fridge at 3 °C for 5 days. The precipitate was filtered off, washed with Me-THF and isopropanol and dried in vacuum oven at 50 °C to give 4.27 g of nearly colourless solid as a mixture of l,2-dideoxy-l,2-di- benzylamino-lactose and benzylammonium hydrochloride.

The ratio of isomeric l,2-dideoxy-l,2-dibenzylamino derivatives in the isolated solid was determined by HPLC : l,2-dideoxy-l,2-dibenzylamino-lactose: 95.2 %; l,2-dideoxy-l,2- dibenzylamino-lactulose : 4.4 %; l,2-dideoxy-l ,2-dibenzylamino-epilactose: 0.4 %. Content of benzylamine hydrochloride was estimated by NMR to be 65 % by weight. Calculated yield was 12 %.

Example 2

A suspension of lactose monohydrate (50 g) and benzylamine (23.1 g, 1.55 eq.) in methanol (100 ml) was stirred at 60 °C for 2 hours followed by slow addition of acetic acid (15.8 g, 1.9 eq.) and the stirring was continued at 60 °C for an additional 2 hours. The obtained solution was co- evaporated with isobutanol giving a suspension to which benzylamine (104 g, 7 eq.) was added and allowed to stir at 30 °C for 15 hours. After addition of benzylammonium hydrochloride (20 g, 1 eq.) the obtained solution was briefly heated up to 45-55 °C to dissolve the solids. The solution was diluted with isobutanol (100 ml) and stirred at 30 °C for 3 days, then further isobutanol was added (100 ml) and the stirring was continued at 40 °C for 7 hours. After addition of methanol (35 ml) the obtained suspension was cooled, the solid was then filtered, washed and dried to give 23.14 g of white material as a mixture of l,2-dideoxy-l ,2-dibenzylaminolactose and benzylamine

hydrochloride. HPLC purity: 99.5 % (CAD), benzylamine hydrochloride assay by HPLC (UV) was 23.3 % by weight. Calculated yield based on purity was 25 %.

Example 3

A suspension of lactose monohydrate (100 g) in benzylamine (190 ml) was stirred at 60 °C for 2.5 hours. The resulting syrup containing N-benzyl-lactosyl amine was co-evaporated with i- BuOH at 60 °C under reduced pressure several times to remove water, then the residue was taken up in methanol (450 ml) and acetic acid (150 ml) was added in portions. The solution was refluxed for 0.5 hours, chilled to ca. 50 °C, and acetone (400 ml) was added. After stirring at 3 °C overnight the precipitated solid was filtered off, washed with acetone and dried to give 110.8 g of 1- benzylamino-l-deoxy-lactulose acetate as a white solid (81.2 %). M.p. : 149-150 °C (dec). HPLC purity: 100 % (CAD detector). NMR (400 MHz, D 2 0, 30 °C) indicated a mixture of three isomers: β-fructopyranosyl (cca. 66 mol%), a-fructofuranosyl (cca. 17 mol%) and β-fructofuranosyl (cca. 17 mol%). X H- and 1 C-NMR assignments are summarized in Table 1.

* Due to extremely overlappiag signals Ή and "C ciiaiical. hifts could not be assigtt«l

Table 1. and 13 C-NMR (400 MHz, D 2 0, 30 °C) assignments of 1-benzylamino-l-deoxy- lactulose acetate salt

Example 4

Lactose monohydrate (10 g) was suspended in isopropanol (50 ml) followed by addition of benzylamine (5 ml) and acetic acid (3.0 ml). The obtained suspension was stirred at 65 °C for 1 hour and at 55 °C for 23 hours, then cooled down, the solid was filtered off, washed with isopropanol (5x10 ml) and dried in vacuo yielding 11.35 g of 1-benzylamino-l-deoxy-lactulose acetate (83 %). Example 5

The suspension of lactose monohydrate (300 g) in methanol (300 ml) and benzylamine (107 g, 1.2 equiv.) was agitated at 65 °C for 1 hour followed by careful addition of acetic acid (75 g, 1.5 equiv ). The obtained mixture was agitated at 65 °C for 6 hours, diluted with iso-butanol and allowed to stir at 3 °C overnight. The title product was isolated by filtration, washed with cold iso- butanol - methanol mixture and dried in vacuum oven at 50 °C to give 337 g of 1 -benzyl amino- 1- deoxy-lactulose acetate as a white solid (83 %).

Example 6

A reactor was loaded with methanol (400 ml), lactose monohydrate (200 g) and benzylamine (122 ml, 2 equiv.) and agitated at 65 °C for 3.5 hours. The temperature of the reaction mixture was lowered to 50 °C then acetic acid (84 ml, 2.5 equiv.) was slowly added under stirring, while keeping the temperature at around 50 °C. The stirring was continued at this temperature for 15 hours. Isobutanol (400 ml) was added and methanol (200 ml) was distilled off under reduced pressure. The reaction mixture was stirred at 0 °C for 1.5 hours before being filtered. The reactor and the filtered solid were washed with a cold isobutanol - methanol mixture and dried in vacuum oven at 50 °C to give 245 g of 1-benzylamino-l-deoxy-lactulose acetate as a white solid (90 %). HPLC purity (relative area) is >99 %.

Example 7

1-Benzylamino-l-deoxy-lactulose acetate (20 g) was suspended in isopropanol (100 ml) with stirring on ice bath and cone. HCl solution (6 ml) was added dropwise. After agitation for 30 min. the solid was filtered off, washed with isopropanol and dried on air. The filtered solid was ground then suspended in hot methanol (40 ml) and diluted with acetone (80 ml), allowed to cool, filtered, washed with acetone and dried in a vacuum oven at 50 °C for 3 hours to give 15.1 g of 1- benzylamino-l-deoxy-lactulose chloride as a colourless solid. M.p.: 166-168 °C (dec). 'H- MR spectrum (D 2 0) was nearly identical to that of the acetate salt, except no acetyl resonance could be found.

Example 8

The suspension of 1-benzylamino-l-deoxy-lactulose acetate salt (80 g), benzylamine hydrochloride (23.4 g, 1 eq.) and benzylamine (92.44 ml) was stirred at 30 °C for 48 hours then isobutanol (80 ml) was added. Stirring continued for another 48 hours followed by addition of isobutanol (80 ml) again. After stirring for an additional 3 days at 30 °C isobutanol (80 mL) was added again and the reaction mixture was cooled to 5 °C to give a suspension containing 1,2- dideoxy-l,2-dibenzylamino-lactose, to which concentrated aqueous HC1 solution (72.13 ml) was added dropwise, keeping the temperature below 10 °C, to reach pH 1-2. After completion of the addition, the reaction mixture was stirred at 5 °C for 30 min. The precipitated benzylammonium chloride was filtered off and the filtrate containing 2-benzylamino-2-deoxy-lactose was used in the next step. Palladium on carbon (10 %, 5.2 g) was added to the filtrate and the pH was adjusted to 3 with 40 % NaOH solution. The reaction mixture was hydrogenated at 8 bar and 45 °C for 30 hours. The catalyst was filtered off and washed with water (20 ml). The aqueous phase was separated followed by slow addition of ethanol (400 ml) at 42 °C in 3 hours Seed crystals were added followed by addition of extra amount of ethanol (400 mL) in 2 hours. The suspension was allowed to cool to +10 °C and stirred overnight. The product was filtered, washed with ethanol and dried in the vacuum oven yielding 13.8 g of lactosamine hydrochloride (22.4 %).

Example 9

The suspension of 1-deoxy-l-benzylamino-lactulose acetic acid salt (10 g), benzylamine hydrochloride (1 equiv ), benzylamine (17.5 g, 8 equiv.) and isobutanol (10 ml) was stirred at 30 °C for 6 days while an additional 10 ml of isobutanol was added after each day. The obtained suspension was allowed to stir at 3 °C for 5 days, filtered, washed with cold isobutanol and dried in vacuo to give 5.4 g of white solid as a mixture of l ,2-dideoxy-l ,2-dibenzylamino-lactose and benzylamine hydrochloride (calculated yield based on NMR purity: 43 %; calculated yield based on HPLC purity: 39 %). The isolated solid (5.3 g) was suspended in water (10 ml) and isobutanol (10 ml) and hydrolysed by pH-controlled drop-wise addition of concentrated aqueous HC1 (2.1 ml) at 5 °C. The cold bath was removed to give a clear solution. To the obtained solution, palladium on carbon (10 %, 0.5 g) was added and hydrogenation was carried out at 10 bar H 2 at room temperature for 18 hours. The catalyst was filtered off. The aqueous phase was separated, the organic phase was washed with water and the combined aqueous phases were concentrated in vacuo to a light brown solid (7.4 g). It was treated with water (1 ml) followed by slow addition of ethanol (19 ml) with sonication. The obtained suspension was stirred at 3 °C for 6 hours and lactosamine hydrochloride monohydrate was isolated by filtration (2.36 g, white solid). The mother liquor gave additional 540 mg and 290 mg in two subsequent crops The total yield was 3 19 g (39.5% for three 3 steps) HPLC (CAD and UV) and NMR purity was 100 %. Example 10

The suspension of 1 -deoxy-l-benzylamino-lactulose acetic acid salt (50 g) was suspended in benzylamine (90 ml, 8 equiv.) and stirred at 30 °C for 24 hours. During this period, azeotropic removal of water was performed with isobutanol. Benzylamine hydrochloride (1 equiv ) and isobutanol (100 ml) were then added and the mixture was stirred at 30 °C for 90 hours, with repeated additions of isobutanol to keep the reaction mixture stirrable (total isobutanol added was 250 ml). MeOH (25 ml) was added before the reaction mixture was cooled to 3 °C and stirred for 22 hours. The solid was filtered, washed with cold iso-butanol (120 mL) then hexane (50 mL), and dried in vacuo to give 29.3 g of white solid as a mixture of l ,2-dideoxy-l,2-dibenzylamino-lactose and benzylamine hydrochloride (calculated yield based on HPLC purity: 43 %). The isolated solid was then hydrolysed with HCl-solution followed by hydrogenolysis as described in example 9 to give 17 g of lactosamine hydrochloride monohydrate (42 % for three steps).

Example 1 1

The protocol according to example 10 was followed with the difference that a) the suspension of 1-deoxy- l -benzylamino-lactulose acetic acid salt and benzylamine was stirred at 50 °C for 5.5 hours, and b) after addition of benzylamine hydrochloride the mixture was stirred at 40 °C for 22 hours. After hydrochloric acid hydrolysis and subsequent hydrogenolysis, 15 6 g of lactosamine hydrochloride monohydrate was obtained (39 % for three steps).

Example 12

The Heyns rearrangement of 1 -benzylamino- 1 -deoxy-D-fructose oxalate salt (the Amadori product of D-mannose or D-glucose with benzylamine, see Ponpipom et al. Carbohydrate Res. 82, 135 (1980) and Levi et al. Bioconj. Chem. 18, 628 (2007)) were performed under two different conditions. The reaction progress was monitored by HPLC and product distribution is summarized in the following table.